From mboxrd@z Thu Jan 1 00:00:00 1970 From: Thierry Bultel Message-ID: <571F77FE.2040100@basystemes.fr> Date: Tue, 26 Apr 2016 16:15:26 +0200 MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8; format=flowed Content-Transfer-Encoding: 7bit Subject: [Xenomai] crash in __pthread_mutex_check_init List-Id: Discussions about the Xenomai project List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: xenomai@xenomai.org, Jean-Baptiste Tredez Hi, I am currently still porting ipipe to kernel 4.5, and after having solved some issues (for instance, I found out that the DTB are no longer compatible, between 3.18.12 and 4.5 and without recompiling the device tree, the hardware address of the timer is wrong, preventing the tsc to work, and making the kernel lockup on whatever sleep call ) ... I wanted to pass the regression tests and I am now facing this one: The posix "leaks" test leads to a kernel crash Apparently, the shadow mutex address returned by _xn_safe_copy_from_user is wrong, and the crash occurs when accessing umx->shadow_mutex.magic I can consider that the problem I have is obviously in my ipipe work, because the same xenomai version works fine with 3.18.12 ... But I do not see any relation for now. I have some questions, please: 1) What is the meaning of CONFIG_XENO_FASTSYNCH ? I see that the _pthread_mutex_check_init syscall only exists when CONFIG_XENO_FASTSYNC is defined (which is the case, here) 2) I have a number of regression tests that pass, proving that the syscall interface is not so completely broken, where should I pay more attention in the ipipe side, in my case ? Thanks Thierry